Beware of Geeks Bearing Gifts: Assessing the Regulatory Response to the Christchurch Call
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
The Political Economy of Communication
7(1), 83–104
© The Author 2019
http://www.polecom.org
Beware of Geeks Bearing Gifts: Assessing the Regulatory
Response to the Christchurch Call
Peter A. Thompson, Victoria University of Wellington, Aotearoa-New Zealand
This commentary is an updated and expanded version of a May 2019 policy
discussion document Beyond the Christchurch Call, prepared for the Better Public
Media Trust (of which the author is board chair).
In May 2019, the Christchurch Call summit in Paris brought together the European Commission, 17
governments and eight major digital media corporations to discuss ways to combat the proliferation
of violent extremist content online. Initiated by New Zealand prime minister, Jacinda Ardern, and
French president, Emmanuel Macron, the summit was impelled by the terrorist attacks on two
mosques in Christchurch in March 2019. Fifty-one people were murdered by a white supremacist
who used GoPro to livestream his rampage on Facebook (literally from a first-person shooter
perspective). The impetus for the Christchurch Call summit stemmed partly from the
unexpectedness of such events transpiring in an ostensibly peaceful and developed country like
New Zealand [1]. However, the use of social media to stream the attacks was a crucial factor in
crystallizing concerns about the harmful consequences of leaving digital intermediaries unregulated.
Although Facebook removed the terrorist video, this came only after they received notification from
the police, by which time the live stream had already finished. This indicated that users were not
reporting the video, and that Facebook’s default algorithms were unable to recognise its nature
(Chang, 2019). In the meantime, 4,000 people had watched the video feed and although Facebook’s
updated algorithms blocked 1.5 million uploads over the ensuing 24 hours, 300,000 versions were
uploaded and had to be removed by moderators (RNZ, 2019a; see also Keall, 2019).
Predictably, the live-feed was also uploaded to a variety of websites tolerant of extremist content
such as 4chan and 8chan, which lie outside New Zealand’s legal jurisdiction and which were
insulated from denial-of-service interference by network services such as Cloudflare (although this
service was withdrawn after the El Paso Walmart massacre - see Taylor and Wong, 2019). The
video was also disseminated through a wide range of mainstream media websites; excerpts
appeared in The Sun, The Daily Mail and The Daily Mirror in the United Kingdom (Waterson,
2019), while in Australia Sky News, channels Seven and Nine and even SBS included some
selected footage in their reports (Williams, 2019). Even a month after the attacks, re-edited versions
with altered digital identifiers were still being discovered on Facebook, Instagram and YouTube,Thompson 84
one of which had 700,000 views (RNZ, 2019b). It was also played at a public rally for Turkish
President, Tayyip Erdoğan.
Gaining support in principle for measures to prevent live-streaming terrorism is hardly
controversial. Nevertheless, the Christchurch Call has opened up a space wherein the issue of social
media regulation can be debated legitimately without the ideological default to neoliberal objections
concerning state interference in the media. This commentary aims to contextualise the initiative
with reference to policy developments in various countries (including the United Kingdom, France,
Australia and New Zealand). It will identify some of the key policy complexities whereby
competing vested interests are seeking to shape or circumscribe regulatory responses, and set out a
broader framework of policy options for addressing issues raised by the activities of social media
and digital intermediaries. It will then argue that although the Christchurch Call opens up an
important space for multi-lateral policy deliberation, the high-level engagement of governments and
corporations risks circumscribing the range of policy options which can be progressed.
Consequently, the Christchurch Call needs to be complemented by robust regulatory responses
which will hold social media and digital intermediaries accountable and defend civic interests
against their incumbent power.
Reluctant self-regulation?
The size and scope of the social media and digital intermediaries coupled with the practical
challenges of brokering international regulatory frameworks has historically persuaded
governments to leave the tech companies unchallenged as the default agents of regulatory change.
Unsurprisingly, the companies have selectively adopted modes of monitoring content while
deflecting attempts to introduce statutory regulations at operational levels where their commercial
interests are liable to be compromised (see Winseck, 2015).
Although Facebook is only one of many social media/ online intermediary companies, it has
become the poster-child for proponents of stronger state-based regulation across the sector. Its role
in enabling the live-streaming of the Christchurch terrorist attacks was preceded by a litany of
ethically dubious practices. More recently, Facebook has been hit by a record US$5 billion fine
from the Federal Trade Commission for privacy breaches, abuses of data and its role in the
Cambridge Analytica scandal (Kelly, 2019b). Facebook certainly has a questionable track record in
blocking extremist content. In 2018, Channel 4’s Dispatches uncovered Facebook’s ‘shielded
review’ moderation system which permitted high-traffic right wing hate speech posted by groups
such as the English Defence League and Britain First to remain online despite numerous
complaints. It was only after the Christchurch terror attack that Facebook acquiesced and blocked
the content in April 2019. Facebook also recently moved to block white nationalist content but as a
Vice Motherboard investigation uncovered, its earlier response to the racially-motivated
Charlottesville murder in 2017 was to block white supremacist content while expediently allowing
white nationalist material to remain (Cox, 2018, 2019; Cox and Koebler, 2019b). Although
Facebook has attracted the most criticism, Twitter and YouTube lag behind it in controlling
extremist content. YouTube algorithms optimize traffic and user consumption by recommending
increasingly extreme content in response to key-word searches. Another recent Motherboard report
identified discrepancies between Twitter’s robust policing of Islamic State content compared with
their lax oversight of right wing content, and cites one executive’s concern that controlling the latter
would result in Republican politicians being blocked (Cox and Koebler, 2019a).Thompson 85
A UK government report was blunt in its assessment of digital intermediaries’ willingness to act
responsibly:
We note that Google can act quickly to remove videos from YouTube when they are
found to infringe copyright rules, but that the same prompt action is not taken when
the material involves hateful or illegal content … The biggest and richest social media
companies are shamefully far from taking sufficient action to tackle illegal and
dangerous content (House of Commons/Home Affairs Committee, 2017, paras
3.1/3.3).
Interestingly, a week before the Christchurch attack, Facebook’s Mark Zuckerberg outlined a
privacy-oriented model for social networking; encrypted messaging tools, data storage and deletion
options for all Facebook apps would give more control over communications to private individuals.
Acknowledging the tension between protecting legitimate dissidents and affording privacy to ‘bad
actors’, he went on to note “a growing concern among some that technology may be centralizing
power in the hands of governments and companies like ours” (Zuckerberg, 2019a). A fortnight after
the attack, however, Zuckerberg published an op-ed in the Washington Post, inviting “a more active
role for governments and regulators” on issues like election integrity (e.g. transparency over
political advertising), privacy, and data portability. He also called for more consistent guidelines on
harmful content, remarking that “Regulation could set baselines for what's prohibited and require
companies to build systems for keeping harmful content to a bare minimum” (Zuckerberg 2019b).
Apart from seeking to redress reputational damage from its facilitation of livestreamed terrorism
and previous abuses of user data (e.g., permitting apps that harvested third party data and
complicity with Cambridge Analytica’s political machinations), Facebook’s new-found openness to
state regulation is founded on self-interest (Cadwalladr, 2017). It is apparent from Zuckerberg’s
remarks that, beyond clearer legal definitions of ‘baselines’, industry actors are still regarded as the
principal agents of regulatory intervention.
Indeed, statutory regulation potentially offers strategic advantages for social media and digital
intermediaries, especially if they are involved in its design and implementation. Media companies
which enable content discovery and sharing between third parties currently operate within an
unspecified regulatory environment between publishing and platform provision. In the absence of
statutory definition of their obligations and liabilities in regard to content management, it is difficult
to assess regulatory risk and the scope or scale of potential penalties, especially when these vary
between jurisdictions.
Formal regulation (such as defining what content is deemed harmful) would allow digital
intermediaries and social media to maintain a default defence of compliance with applicable laws
(which could also specify the maximum penalties for violations) (see Isaac, 2019). As the pro-free
market Mises Institute observes, increasing compliance costs (e.g., requiring social media to expand
content monitoring) favours market incumbents like Facebook. Moreover, “By offloading decisions
about harmful content, privacy rules, and elections onto third-parties, Facebook may not have to
take as much of the heat when mistakes are made” (McMaken, 2019).
The return of the regulators?
The political momentum toward regulation of online/social media and other digital intermediaries
was already well under way well before the terrorist attack in Christchurch and the subsequentThompson 86
summit in Paris. Unsurprisingly, in the wake of the Christchurch attack, the resolve of governments
to regulate social media and digital intermediaries has hardened, with a variety of responses. As
Annany and Gillespie (2017) and Flew, Martin and Suzor (2019) have presciently observed, policy
responses rushed through in response to ‘public shocks’ such as mass shootings risk being poorly
designed and may deliver unintended outcomes. Indeed, there is a risk that policymakers under
public pressure to take action on social media will focus on immediate symptomatic issues and
misdiagnose deeper structural problems.
Some researchers and policy actors have sought to designate social media and digital
intermediaries as media content publishers in order to bring them under existing regulatory
frameworks, just as the corporations have often resisted such definitions (Napoli and Caplan, 2017).
As New Zealand’s prime minister, Jacinda Ardern recently observed, “We cannot simply sit back
and accept that these platforms just exist and that what is said on them is not the responsibility of
the place where they are published. They are the publisher. Not just the postman.” (quoted in Small,
2019). Others have argued that defining digital intermediaries as media publishers invokes the
wrong regulatory paradigm for many of the contemporary policy concerns. Winseck (2019), for
example, suggests that the definition of content in these circumstances cannot be equated to the
content of editorial work or publishing, and cautions that extremist content online can often be
addressed adequately by existing laws. This view does not assume that digital intermediaries and
platform providers are neutral channels of transmission with no content-related responsibilities.
Their market scale and power certainly needs to be addressed through statutory measures. Winseck
(2019) suggests functionally equivalent data-privacy protection measures across all layers of the
value chain, including information fiduciary [2] obligations on platforms/intermediaries as well as
requirements for algorithmic transparency and audits.
In the United States, senator Elizabeth Warren has called for anti-trust measures to break up
dominant ‘platform utilities’ such as Amazon, Facebook and Google, arguing that they exert “ too
much power over our economy, our society, and our democracy. They've bulldozed competition,
used our private information for profit … [and] hurt small businesses and stifled innovation”
(quoted in Caplin, 2019). The European Commission, in conjunction with several major tech firms,
introduced a Code of Conduct in 2016 to restrict the proliferation of online hate-speech. Germany,
meanwhile, introduced the 2017 Network Enforcement Act (NetzDG) which requires the removal
of hate speech within 24 hours and imposes fines for failing to respond (in fact Facebook now
employs almost a sixth of its global content moderation staff in Germany) (see Flew et al., 2019).
However, opponents have argued that NetzDG has led moderators to block contentious but
legitimate expression (Kinstler, 2019).
In the United Kingdom, the 2017 House of Commons, Home Affairs Committee report on
online hate speech and extremism identified the reliance of social media on the user community to
police content as inadequate. More robust measures were needed because “the interpretation and
implementation of the community standards in practice is too often slow and haphazard” (2017,
para 3.9). The report also noted the inapplicability of traditional frameworks of media regulation to
social media and digital platform operators.
In 2018, the government also announced the introduction (from 2020) of a 2 percent levy on the
domestic turnover of digital intermediaries with global revenue of over £500m (BBC 2019a).
Intended primarily as a remedial initiative to force global tech companies to pay the tax they
otherwise avoid by declaring profits in offshore havens, the initiative is important because it
reclaims online commercial turnover as domestic economic activity. The recent Online HarmsThompson 87
White Paper (Department for Digital, Culture, Media and Sport/ Home Office, 2019) sets out a
series of “duty of care” responsibilities for digital intermediaries under a new regulatory body with
the power to impose fines for violations. The onus would be on social media companies to limit
harmful content (notably extremist/terrorist material and child sexual abuse). The duty of care also
extends to increasing source transparency and reducing the propensity for filter-bubbles to
proliferate disinformation (see sections 7.27-7.31).
France has also been pro-active in its response to social media regulation. President Emmanuel
Macron, along with Jacinda Ardern, was, of course, instrumental in bringing state and industry
together for the Christchurch Call summit in Paris. In May 2019, an interim report (commissioned
before the Christchurch attacks) setting out the French framework for social media regulation was
published. The regulatory team was granted six months of access to Facebook (France 24, 2019)
and concluded that:
Even if the abuses are committed by users, social networks’ role in the presentation
and selective promotion of content, the inadequacy of their moderation systems and
the lack of transparency of their platforms’ operation justify intervention by the
public authorities, notwithstanding the efforts made by certain operators (Office of
the Secretary of State for Digital Affairs, 2019: 10).
The report recommended a range of regulatory measures, including an “independent administrative
authority, acting in partnership with other branches of the state, and open to civil society” (3). Three
key transparency obligations were also cited: algorithmic ordering of content; terms of service and
content moderation; and a requirement to defend user integrity (equivalent to a ‘duty of care’) (21).
In July 2019, the French Parliament also approved a bill obliging digital intermediaries to remove
content deemed to promote ‘obviously hateful’ speech. This included extremist, violent, racial and
religious discrimination and child pornography with fines of up to €1.25 million if the content was
not blocked within 24 hours (Kelly, 2019a). However, this initiative is subject to senate approval
and some critics have suggested that it gives too much control over content to the intermediaries
(The Guardian, 2019).
In a separate initiative, the government announced a 3 percent levy on the domestic turnover of
major tech corporations with annual (global) turnover of over €750m. The specific targets were
digital advertising and the sale of personal data to facilitate advertising. As with the United
Kingdom’s digital tax model, this effectively pre-empts a very similar European Commission
proposal advanced earlier in 2019 (BBC, 2019a, 2019b). There has, thus far, been a highly negative
response to these developments from the Trump administration because most of the digital
intermediaries falling under the new regime are based in the United States.
In Australia, there was a swift legislative reaction to the Christchurch terror attack: The new
Unlawful Showing of Abhorrent Violent Material Bill took barely a fortnight to be approved by
Parliament (Attorney-General for Australia, 2019). Social media companies failing to remove
extreme content (e.g., terrorism, murder and rape) in a timely manner after notification from a new
e-Safety commissioner would be liable either for a fine up to 10% of their turnover or a sentence of
three years imprisonment for the responsible executives. Minister for Communications, Mitch
Fifield, explained the move thus:
Mainstream media cannot live broadcast the horror of Christchurch or other violent
crimes and neither should social media be able to do so … Where social mediaThompson 88
platforms fail to take action to stop the live streaming of such violent and abhorrent
crimes, they should face serious penalties and that's what will now occur once this Bill
receives Royal Assent (quoted in Attorney General for Australia, 2019).
Although the legislation targeted a narrow range of particularly extreme content, the Digital
Industry Group Inc. (representing Facebook, Google and Twitter) has argued that the bill was
rushed through with insufficient deliberation and that United States law prevented these groups
from sharing content data (Bogle, 2019). Three months later, the report of the Australian
Competition & Consumer Commission (ACCC) Digital Platforms Inquiry (July 2019) was
published, recommending a wider range of measures to redress the market power of digital
intermediaries. These include developing a harmonized regulatory framework for digital media on a
platform-neutral basis alongside a code of conduct to promote a fairer balance of relations between
digital intermediaries/platforms and news media. Flew (2019) suggests that the ACCC’s platform-
neutral approach is important, as is its recognition that digital intermediaries require specific
measures that currently fall between competition and platform-neutrality measures and Australian
Communications and Media Authority’s (ACMA) content regulation regimes. The ACCC report
also notes the need for large digital platform providers with over a million users to develop a code
of practice to address disinformation and fake news. However, unlike the Abhorrent Violent
Material Bill, this appears to place the onus back on the digital intermediaries to develop their own
criteria for managing online content. The proposals also contain measures to strengthen privacy
protections, including consumer consent for data sharing, more rights for consumers to opt out
and/or seek legal redress for breaches.
In Aotearoa-New Zealand, the complexities of digital convergence have been recognised by
successive governments, but to date there has been no fundamental overhaul of the regulatory
framework since the neoliberal reforms of the late 1980s (although the 2008 Review of Regulation
initiative attempted this before being canned) (Thompson, 2019). The 2015 Exploring Digital
Convergence consultation (Ministry for Culture and Heritage/ Ministry of Business Innovation and
Employment, 2015) sought to resolve the gaps in the Telecommunications and Broadcasting Acts
(including online services and content standards) but the ensuing legislation never progressed. In
New Zealand, it is apparent that policy initiatives addressing the issues surrounding social media
and digital intermediaries (along with other communication policy issues) have become somewhat
fragmented across different government departments. As Mason and Errington (2019) observe, the
regulation of social media is “being addressed in a piecemeal fashion by an array of government
agencies, including the Privacy Commission, the Ministry of Justice, the Department of Internal
Affairs, and Netsafe” (2019:4).
Nevertheless, the 2015 Harmful Digital Communications Act (see Ministry of Justice, 2017),
does address cyber-bullying and harassment, with measures to restrict harmful, threatening and
offensive messages. However, the framework mainly targets peer-to-peer abuse and does not
encompass broader issues concerning hate speech facilitated by digital intermediaries and social
media.
The immediate regulatory response to the Christchurch attack was a move by the chief censor to
classify the terrorist’s live-streamed video and accompanying manifesto as “objectionable”. The
penalties for possession or distribution are fines of up to NZ$10,000 or imprisonment of up to 14
years (Department of Internal Affairs/Te Tari Taiwhenua, 2019a). Thus far, one neo-nazi has been
jailed for 21 months for illegally distributing the video.Thompson 89
A Royal Commission of Inquiry into the attack on the Christchurch mosques was also announced in
April (Department of Internal Affairs/Te Tari Taiwhenua, 2019b). This is focused primarily on
investigating why state agencies were unable to anticipate or prevent the Christchurch attack, but it
is likely that the use of social media by extremists (and police monitoring thereof) will be
highlighted. It is also significant that, in April 2019, the government pushed through a bill banning
military-style semi-automatic firearms and moved to set up a buy-back scheme despite protestations
from the local gun lobby (a legal initiative that still seems unthinkable in the United States
notwithstanding the recent El Paso and Dayton mass shootings).
Several New Zealand-based NGOs and think tanks also published extensive commentaries and
reports in advance of the Christchurch Call. InternetNZ, facilitated a “civil society” statement on the
Christchurch Call response informed by the Voices for Action meeting in Paris the day before the
Paris Summit (InternetNZ, 2019). Focusing on human rights issues, the report argues for a free and
open internet as core principles. Although supportive of regulation on terrorist and other extremist
content, the report highlights the need to define such terms carefully and avoid inadvertently
inviting authoritarian responses which could harm civic interests:
It is of vital importance that governments participating in the Christchurch Call
commit to robust accountability and oversight to ensure that laws, mechanisms, and
other initiatives to combat terrorism online do not result in disproportionate human
rights violations of political critics, human rights defenders, journalists, ethnic or
religious minorities, refugees, asylum seekers, and migrants (InternetNZ, 2019: 1)
The report also emphasised the need to differentiate between social media and infrastructure
providers and insisted that regulatory measures not impinge on open access to the infrastructures of
the internet (see also Carter and Komaitis, 2019). The Helen Clark Foundation (Mason and
Errington, 2019) noted that, even if the social media companies did not intend to promote extremist
content per se, they are highly motivated to optimise traffic through the sharing and discussion of
controversial material. Cautioning against a regulatory default to the United States framework of
media regulation, the report supports a new regulatory body along with a statutory “duty of care”
for social media platforms (including reasonable measures to develop technologies and mechanisms
to minimise harm). Another report from the Workshop (Elliott, Berentson-Shaw, Kuehn and Salter,
2019) identified three overarching regulatory challenges: platform monopolies; algorithmic opacity;
and the constraints of the attention economy. The report supports regulating digital intermediaries
and social media to control extremist content, but stresses the need for clear definitions.
Importantly, it highlights the need for a response that extends beyond content regulation to consider
a wider range of structural issues in the digital media ecology. To this end, it also calls for
technological solutions in the design of digital platform architectures to encourage civic
participation.
The Christchurch Call: tip of the iceberg?
New Zealand’s prime minister, Jacinda Ardern, gained international attention and praise after her
empathetic support for the Christchurch Muslim community in the aftermath of the attacks. Images
of her wearing a hijab and hugging one of the survivors were published in news media around the
globe. Jacinda Ardern, working alongside the French president, Emmanuel Macron, attracted global
media attention and provided a persuasive pretext for government and industry to attend theThompson 90
Christchurch Call summit. In addition to the European Commission, government actors included
Australia, Canada, France, Germany, Indonesia, India, Ireland, Italy, Japan, Jordan, the
Netherlands, New Zealand, Norway, Senegal, Spain, Sweden and the United Kingdom (the United
States was conspicuous by its absence, although a White House statement later endorsed the anti-
terrorist aims of the Christchurch Call - see White House, 2019). Meanwhile, the industry actors
present were Amazon, Daily Motion (Vivendi), Facebook, Google, Microsoft, Qwant (French
search engine provider), Twitter and YouTube (also a Google Subsidiary).
The ensuing pledge document (the Christchurch Call, 2019) is only three pages long and non-
binding. It is obviously unrealistic to expect more from a one-day summit. Indeed, the achievement
of getting government and key industry actors to sit together at the table and agree on basic
principles arguably matters more than the content of the ensuing pledge, insofar as it sets the stage
for future deliberations and multilateral engagement among government actors and industry.
However, the fact that civic actors were left to meet separately is perhaps indicative of the narrow
scope of debate. Had the Paris summit failed to produce any sort of multilateral agreement, it might
well have backfired and served to reinforce the perception that the global tech firms cannot be
effectively regulated. The document outlines broad commitments and principles in three areas, the
edited highlights of which are listed below (the Christchurch Call, 2019):
Governments
• Counter the drivers of terrorism and violent extremism by strengthening the resilience and
inclusiveness of our societies.
• Ensure effective enforcement of applicable laws that prohibit the production or
dissemination of terrorist and violent extremist content.
• Encourage media outlets to apply ethical standards when depicting terrorist events online.
• Support frameworks, such as industry standards, to ensure that reporting on terrorist attacks
does not amplify terrorist and violent extremist content.
• Consider appropriate action to prevent the use of online services for the purposes of
disseminating terrorist and violent extremist content.
Online Service Providers
• Take transparent, specific measures to prevent the upload of terrorist and violent extremist
content and its dissemination on social media and similar content-sharing services.
• Provide greater transparency in the setting of community standards or terms of service.
• Enforce those community standards or terms of service in a manner consistent with human
rights and fundamental freedoms.
• Implement immediate, effective measures to mitigate the specific risk that terrorist and
violent extremist content might be disseminated through livestreaming.
• Implement regular and transparent public reporting.
• Review the operation of algorithms and other processes that may drive users towards
terrorist and violent extremist content in order to identify possible intervention points.
• Work together to ensure cross-industry efforts are coordinated and robust.Thompson 91
Governments and online service providers
• Work with civil society to promote community-led efforts to counter violent extremism in
all its forms.
• Develop effective interventions based on trusted information sharing in regard to
algorithmic and other processes (so as to redirect users away from terrorist and violent
extremist content).
• Accelerate research into the development of technical solutions which will detect and
immediately remove terrorist and violent extremist content online.
• Support research and academic efforts to better understand, prevent and counter terrorist and
violent extremist content online.
• Ensure appropriate cooperation with and among law enforcement agencies for the purposes
of investigating and prosecuting illegal online activity in regard to terrorist and violent
extremist content.
• Support smaller platforms as they build capacity to remove terrorist and violent extremist
content.
• Collaborate with, and support partner countries in the development and implementation of
best practice in preventing the dissemination of terrorist and violent extremist content
online.
• Develop processes allowing governments and online service providers to respond rapidly,
effectively and in a coordinated manner to any major event which involves the
dissemination of terrorist or violent extremist content.
• Avoid directly or indirectly contributing to adverse human rights impacts through their
business activities and by addressing such impacts as and when they occur.
• Recognise the important role of civil society concerning the issues and commitments in the
Christchurch Call.
The predominant focus on measures curtailing terrorist and extremist content, while understandable
in the short-term, also underlines the narrow premises of agreement among government and tech
sector actors. The cursory acknowledgement of human rights and the role of civil society do
nothing to guarantee that future deliberations will not be quarantined within a narrow content-
oriented framework.
There are three important arguments for pursuing a more extensive approach to the regulation of
social media and digital intermediaries, both within domestic jurisdictions as well as through the
multilateral Christchurch Call framework.
1. The concerns over online hate-speech and extremism, while important, are symptomatic of
deeper structural patterns in the digital media ecology, including:
a) Intensified financialisation and commercial competition (with commensurate increases
in market failures and the opportunity costs of maintaining civic and cultural
obligations).
b) Disruption of traditional value chains and business models through convergence.
c) New forms of network dominance stemming from digital intermediaries’ control over
the architectures and algorithms of content discovery (and thereby audience traffic and
associated revenues).Thompson 92
2. The response to the Christchurch terror attack and the ensuing Christchurch Call summit has
opened up a space of policy deliberation wherein proposals to regulate social media and digital
intermediaries and hold them accountable to the public interest can be tabled without inviting
preemption of such initiatives. Such a response could take the form of wholesale opposition
either from vested commercial interests within the tech sector or from governments and state
departments sceptical of regulatory intervention in markets.
3. Although the Christchurch Call framework has brought several governments and key industry
actors to the table, the domestic regulatory responses to the growing power of social media and
digital intermediaries in respective domestic jurisdictions remain vital because:
a) The probability of substantial, binding, multi-lateral agreements on social media/digital
intermediary regulation in the short term is negligible. The prospect of developing such
a framework beyond protocols for removing extremist/terrorist content is, at best,
uncertain. Delaying domestic interventions on the pretext that the global tech companies
can only be meaningfully regulated through a multilateral accord risks further
entrenching their incumbent dominance. Future claw-backs to make them accountable to
civic interests will be even more difficult.
b) Many forms of regulatory intervention that might ensue from a multilateral forum will
still need to be implemented and enforced on a domestic level. The scope of the
possibilities for intervention at a multilateral level are likely to be informed by working
examples of existing interventions (as well as associated lobbying from the global tech
companies).
c) It is likely that the threat of various regulatory measures implemented through different
state agencies reflecting different policy rationales has compelled the global tech
companies to sit down at the bargaining table. They face the prospect of levies on
turnover to redress tax avoidance with fines for inadequate monitoring of extremist
content and duty of care obligations to protect user privacy. These companies would
rather negotiate a consistent framework at a multilateral forum than confront a multitude
of regulatory measures which might compromise different parts of their business model
in different countries.
A framework for policy deliberations
The scope of this discussion does not extend to specific policy prescriptions, not least because their
shape and form needs to be articulated within the institutional arrangements associated with the
media ecologies of each jurisdiction. I will, however, outline a heuristic framework for working
through the regulatory issues and potential policy responses. To this end, a ‘value chain’ model is
set out, identifying different layers of the media sector which represent potential points of
intervention. Importantly, this differentiates between the levels of content distribution and content
discovery. This is a crucial point where the platforms and algorithms of social media and other
digital intermediaries have become dominant (see Figure 1).Thompson 93
Figure 1. A value-chain model of regulatory intervention pointsThompson 94
One important implication of the model here is that ‘upstream’ levels can affect ‘downstream’
levels (and in some cases vice-versa). Thus, the shape of ownership and level of market
concentration will affect content priorities (such as driving news providers to seek online traffic
through ‘clickbait’). New forms of content discovery via mobile/online platforms may fragment
audiences and influence content formats and distribution channels (e.g., ‘digital first’ policies).
Concomitantly, regulatory interventions at one level may have implications (whether intentional
or inadvertent) for others. For example, imposing copyright restrictions on online content sharing or
imposing a levy on online advertising spending would influence the business models of digital
intermediaries. Similarly, subsidising the expansion of broadband networks affects consumer take-
up of on-demand content streaming services. Obviously, this model does not make predictions
about the efficacy of specific policy interventions. That depends on the configuration of the media
market and the institutional priorities of the media actors comprising it. However, the model does
invite a more holistic approach to identifying regulatory interventions.
Applying the framework to social media after the Christchurch Call
Even if there is a broad consensus that something needs to be done about extremist, terrorist and
hate speech online, generic calls to ‘regulate Facebook’ and other social media are, in themselves,
unhelpful. The nature and source of the problems must be correctly identified and regulatory
interventions targeted appropriately and proportionately. Four key questions arise here:
1) What exactly is the issue to be resolved and what policy outcome is desired?
2) At what point of the value chain should regulatory intervention be focused?
3) Which agents are responsible for implementing and enforcing regulatory
measures?
4) What mechanisms for delivering the policy outcomes are to be employed?
A table outlining some of the policy issues, points of intervention, agents of regulation and potential
mechanisms on social media issues is presented in Figure 2. Preliminary discussion is needed to
illustrate some of the technical and normative tensions which are liable to arise.
As noted earlier, the Christchurch Call’s primary focus is to control the dissemination of
terrorist/extremist content through online media and there appears to be a fundamental consensus
that regulatory intervention is needed. However, the point at which content could or should be
subject to regulation can be debated, especially if the ultimate policy outcome of the Christchurch
Call is to reduce terrorist activity itself, not just its livestreaming. A major tension arises here: if law
enforcement agencies were permitted access to social media data to pre-emptively identify potential
terrorists and extremists, this would inevitably infringe personal privacy and civil liberties (while
extremists still inhabited the ‘dark web’).
Facebook’s call for regulatory guidance on its content moderation also reflects the intrinsic
complexity of deciding what criteria can be incorporated into algorithms or moderator practices.
Facebook’s decision to block groups like Britain First is arguably justified on the basis that they
actively promoted racial and religious intolerance. However, the United Kingdom government does
not currently class Britain First as a banned organisation (even though this is the case for militant
far right group National Action). One might therefore contend that unless all the views expressed
by people affiliated with such groups were intrinsically objectionable, free speech is imperiled when
popular sentiment or political correctness drives social media companies to impose wholesale bansThompson 95
on otherwise legal public organisations. The question arises, who should arbitrate the views that are
permitted online?
As noted earlier, the willingness of social media companies to comply with official demands for
content removal has often been limited. If self-regulatory measures are insufficient then regulators
could require internet service providers (ISPs) to block access to websites designated as
objectionable. However, controlling content discovery by restricting internet access is opposed by
some civic groups as a step toward controlling the infrastructure and as a threat to open access.
Technically speaking, it is not unduly difficult and can target specific URL addresses. The risk is
that such measures could become a heavy-handed default in lieu of more nuanced policy measures.
Of course, countries with authoritarian media policy frameworks such as China have gone even
further by restricting public access to entire platforms such as Facebook and Google. Singapore
recently introduced legislation to control disinformation which gives the government the power to
order the blocking or retraction of content that it deems to be fake news. In effect, therefore,
Singaporean government agencies determine what is or isn’t true (see Fullerton, 2019).
Another complexity in policing extremist content is that it is easier to respond retro-actively
than pre-emptively. The measures needed to pre-emptively stop the livestreaming of actions such as
the Christchurch attacks could impinge on other democratic freedoms. For instance, the pre-vetting
and registration of all livestream users would not necessarily exclude a heretofore unknown
terrorist, but could be used to stop known democracy activists from livestreaming a protest. One
could improve regulatory oversight of the algorithms used by social media and online
intermediaries to harvest user data and prioritise content in social media feeds or search engines.
This would, however, require an unprecedented level of transparency from a sector reluctant to
allow outside scrutiny. Analysing the code to identify those factors liable to trigger
recommendations of extremist content is not simple, but it might ensure that such material was not
easily discovered by users. The possibility of governments deciding what kind of content should be
deprioritized could raise democratic concerns. Moreover, a wholesale restriction on algorithms
designed to magnify the visibility and discoverability of ‘trending content’ would affect the
business models of digital intermediaries.
Calls for the regulation of social media and digital intermediaries often assume that they are
more than just platform providers but content aggregators, curators, distributors and/or publishers.
The implied dichotomy between media companies which produce and publish content and those
which provide the ‘pipes’ for the common carriage of other’s content is often unhelpful for the
understanding of digital media ecology. However, the primary operations of search engines and
social media and other digital intermediary services do not entail provision of either the distribution
infrastructure or the content forms themselves. Rather, they provide the intermediary architecture
and navigation platforms which enable third parties to share and discover content (in order that they
can monetise the online traffic thereby generated). This is also why the core operations of social
media and search engines tend to fall between the traditional regulatory provisions for
telecommunications and broadcasting. Further, this suggests that the prevailing regulatory
frameworks need more than cosmetic reform to address concerns related to the practices of social
media and digital intermediaries.
Table 1 presents a breakdown of several regulatory issues arising from concerns about social
media and digital intermediaries. Potential responses are categorized in terms of the point of
intervention on the value-chain, possible agents of intervention and the potential mechanism for
addressing the issue. The identification of certain regulatory agents and mechanisms is intended toThompson 96 offer a framework for considering a broader potential regulatory response, not to endorse any option as desirable or feasible.
Thompson 97
Table 1. Value chain model applied to social media and digital intermediaries in the Christchurch Call context
Potential points of Possible agents of
Regulatory issues Potential mechanisms/modes of intervention
intervention intervention
Proliferation of harmful extremist • Content production • Police/law enforcement • Prosecution of terrorist/extremist groups
content including online hate-speech • Content form (domestic or multilateral) • Increased policing of ‘dark web’
and live-streaming of terrorist acts • Distribution platforms • Content standards regulators • Pre-vetting, classification & restriction of content pre-
(e.g. Christchurch mosque attacks) • Advertisers distribution to resrict objectionable material
• Content discovery & algorithms
• Audience reception • Internet Service Providers • Pre-vetting for live-streaming access
• Digital Intermediaries (social • Development of codes of online practice for digital
media, search engines) intermediaries
• Algorithms • Content standards codes vetted post-distribution in response
• Audience to complaints and/or via algorithms
• Content providers self-police content or regulators issue
take-down notices
• ISP website blocking
• Advertiser boycotts of non-compliant social media (or
possible restrictions on advertising)
• Algorithmic oversight & accountability
• Audience media literacy & self-monitoring, also complaints to
regulators and authorities, consumer boycotts
Algorithms underpinning search • Distribution platforms • Digital intermediaries • Increased self-monitoring of algorithms by social media and
engines and social media news-feeds • Content discovery/algorithms • Independent regulators search engines
promote extremist echo chambers & • Audience reception • Government • Independent regulator access to and vetting/ oversight of
filter-bubbles), undermining rational algorithms/code (e.g. requiring prioritisation of independent
• Audience
dialogue/ community solidarity. public interest media in news-feeds and web searches)
• Government support for public interest media inlcuding
public service broadcasters and independent journalism
• Audience media literacy
Collation and (a)buses of personal • Content licensing/IP • Police/law enforcement • Legislation to protect privacy of audience personal data as
data/ breaches of privacy either by • Content discovery/reception • Independent regulators inalienable intellectual property.
social media or third parties to whom • Algorithms • Internet Service Providers • Require active consent for data sharing with third parties and
data is made available (e.g. Facebook rights to revoke consent and delete stored data.
• Audience reception • Digital Intermediaries (social
traded user data to Amazon in return • Imposition of information fiduciary duty of care on social
media, search engines) plus
for advertising business) media companies’ use of personal data
associated ‘App’ providers
• Audience • Permit law enforcement agencies to access personal data in
order to anticipate extremist/terrorist activity
• Enhanced audience rights to block undesired advertising
• Audience media literacy, consumer boycottsThompson 98
Potential points of Possible agents of
Regulatory issues Potential mechanisms/modes of intervention
intervention intervention
Co-option of social media (and/or • Distribution platforms • Police/law enforcement • Legislation to criminalise concerted efforts to manipulate
search engines) and user data by • Content discovery • Independent regulators elections through abuse of social media
political agencies for targeted • Audience reception • Digital Intermediaries (social • Expanded fact-checking and fake news detection/removal
dissemination of propaganda and/or media, search engines) systems by social media and/or regulators
fake news (e.g. Cambridge Analytica ) • Blocking of identified fake news proliferators from social
• Internet Service Providers
• Audience media and other online platforms
• Audience media literacy
Impact of digital intermediaries on • Content licensing/IP • Government • Regulation obliging social media and search engines which
traditional media (e.g. newspapers) by • Distribution platforms • Independent regulator facilitate third party audience discovery to contribute to the
dominating advertising revenue from • Content discovery/reception • Content producers cost of producing the content thereby discovered/shared.
online traffic through facilitating • Impose a marginal levy on domestic advertising spend
• Audience reception • Digital Intermediaries
discovery of third party content (e.g. directed toward domestic audiences at point of transaction,
it is estimated that Google and • Advertisers
with the revenue redirected to content providers.
Facebook account for 70% of online • Subject existing relationships between digital intermediaries,
advertising) content providers and audiences to market competition/
market power rules.
Concentration, Network effects & • Media institutions • Government (multi-lateral/ • Redesignate digital intermediaries as public utilities with civic
monopolisation of the means of • Content discovery/reception domestic) obligations beyond private shareholders.
online content discovery (e.g. it is • Distribution platforms • Independent regulator • Subject existing relationships between digital intermediaries,
estimated that Google and Facebook content providers and advertisers to market competition/
account for 80% of online referrals) market power rules.
• Require formal break-up or partial nationalisation of digital
intermediaries which exert disproportionate market power
or fail to operate in the public interest.
• Government support for public interest media inlcuding
public service broadcasters and independent journalism
Concentration, Network effects & • Media institutions • Government (multi-lateral/ • Redesignate online-e-commerce providers as public utilities
monopolisation of the platforms for • Distribution Platforms domestic) with civic obligations beyond private shareholders.
e-commerce. • Independent regulator • Subject e-commerce platform providers to market
competition/ market power rules.
• Require formal break-up or partial nationalisation of e-
commerce providers which exert disproportionate market
power or fail to operate in the public interest.Thompson 99
Concluding points
The Christchurch Call pledge represents a very initial step toward the formation of a multilateral
regulatory framework for controlling online terrorist and extremsit content, along with other
practices of social media and online intermediary operators. It would be premature to declare the
initiative either a success or a failure at this stage, but it needs to be understood in the context of a
broader policy trajectory which has been unfolding over the past decade. Many of the regulatory
initiatives being implemented were either in the policy pipeline or under deliberation well before
the Christchurch terror attacks and the ensuing Paris summit. It would, therefore, be wrong to
characterise these measures as being driven by the shock of these events (although the Australian
Unlawful Showing of Abhorrent Violent Material legislation may be the exception).
The Christchurch Call proposal itself is significant insofar as it brought a range of state and
industry actors together and managed to find sufficient common ground for developing a
multilateral, mutli-stakeholder agreement on future principles and responsibilities. As such, it
provides a basis for progressing deliberations on regulatory measures for social media and digital
intermediaries in the future. It is important to bear in mind that the Christchurch Call pledge
document is not sufficiently specific to be enforceable, even if it were enacted into some form of
legally-binding treaty. It also makes only cursory reference to broader questions of human rights
and civic accountability, the regulatory import of which extends well beyond the immediate
question of curtailing online extremism. What the Christchurch Call does do though is legitimate
future regulatory interventions in the digital media sector. They can no longer be casually dismissed
by vested interests or ideological opponents of state intervention in markets.
A consistent, international multilateral framework of regulation must be the ultimate goal of the
Christchurch Call, but the corporate interests must not be be allowed to circumscribe its scope. The
parallel domestic regulatory initiatives stemming simultaneously from different government
departments in different jurisdiction may be somewhat disparate, but deferring domestic level
regulation in the hope of an imminent multilateral regulatory framework would be both politically
naïve and risky. Indeed, the groundswell of support for statutory regulation of social media and
digital intermediaries in different jurisdictions has probably helped to bring the global tach
companies around the bargaining table.
Meanwhile, one must be cautious of the new-found willingness of social media corporations and
digital intermediaries to engage with governments and regulators after years of evading and denying
responsibility or liability. Indeed, the strategic interests of these corporations may actually be served
by a loose, fragmented international framework in which they remain the primary regulatory agents
of intervention. Such an outcome is likely if social media corporates are able to obstruct or narrow
the call for clearer legal guidelines on industry self-moderation of extremist content. We should
beware of geeks bearing gifts.
Author bio
Peter Thompson is a senior lecturer in the Media Studies programme at Victoria University of
Wellington. In addition to co-editing the Political Economy of Communication journal, he is co-
vice-chair of the Political Economy section of the International Association for Media and
Communication Researchers (IAMCR). He has published extensively on media policy in Aotearoa-
New Zealand and is chair of the Better Public Media trust.Thompson 100
Endnotes
[1] It is worth noting that the Ogassagou and Welingara massacre of 160 Futani
herders in Mali, just one week after the Christchurch attack, received very little
western media attention. Although the Easter bombings in Sri Lanka the
following April which claimed 259 lives did gain considerable coverage, its
profile as a news story reflected the fact that tourist hotels were a primary
target.
[2] The notion of ‘information fiduciaries’ suggests an obligation like a financial
or legal professional’s duty of care to protect personal data and use it only to
serve the interests of the client- see Balkin and Zittrain (2016)
References
ACCC (2019) Digital Platforms Inquiry: Final report. Available at:
https://www.accc.gov.au/system/files/Digital%20platforms%20inquiry%20-%20final%20report.pdf
(accessed July 27 2019).
Annany M and Gillespie T (2017) Public platforms: Beyond the cycle of shocks and exceptions. Paper
presented to the 67th Annual Conference of the International Communications Association,
Interventions: Communication Research and Practice. San Diego, 25–29 May. Available at:
http://blogs.oii.ox.ac.uk/ipp-conference/sites/ipp/files/documents/anannyGillespie-publicPlatforms-
oii-submittedSept8.pdf (accessed 24 June 2019).
Attorney General for Australia (2019) Tough new laws to protect Australians from live-streaming of violent
crimes. Government press release, 4 April. Available at:
https://www.attorneygeneral.gov.au/Media/Pages/Tough-New-Laws-to-protect-Australians-from-
Live-Streaming-of-Violent-Crimes.aspx (accessed 27 June 2019).
Balkin JM and Zittrain J (2016) A grand bargain to make tech companies trustworthy. The Atlantic, 3
October. Available at: https://www.theatlantic.com/technology/archive/2016/10/information-
fiduciary/502346/ (accessed 23 June 2019).
BBC (2019a) France tech tax: What's being done to make internet giants pay more? BBC Business, 11 July.
Available at: https://www.bbc.com/news/business-48928782 (accessed 13 July 2019).
BBC (2019b) France passes tax on tech giants despite US threats. BBC World News, 11 July. Available at:
https://www.bbc.com/news/world-europe-48947922 (accessed 13 July 2019).
Bogle A (2019) Laws targeting terror videos on Facebook and YouTube 'rushed' and 'knee-jerk', lawyers and
tech industry say. ABC News, 4 April. Available at: https://www.abc.net.au/news/science/2019-04-
04/facebook-youtube-social-media-laws-rushed-and-flawed-critics-say/10965812 (accessed 24 June
2019).
Cadwalladr C (2017) The great British Brexit robbery: how our democracy was hijacked. The Guardian, 7
May. Available at: https://www.theguardian.com/technology/2017/may/07/the-great-british-brexit-
robbery-hijacked-democracy (accessed 25 June 2019).
Carter, J. and Komaitis, K. (2019) How to regulate the internet without shackling its creativity. Stuff, May
15. Available at: https://www.stuff.co.nz/national/christchurch-shooting/112704241/how-to-
regulate-the-internet-without-shackling-its-creativity (accessed 20 June 2019).You can also read
